Name: Introducing Logikcull ASK in Europe: Pragmatic Legal AI That Gets You to the Heart of the Matter Fast
Uploaded: 2026-03-19T16:16:13.001Z
Duration: 31 min 8 s
Description: Introducing Logikcull ASK in Europe: Pragmatic Legal AI That Gets You to the Heart of the Matter Fast

Transcript for "Introducing Logikcull ASK in Europe: Pragmatic Legal AI That Gets You to the Heart of the Matter Fast": Alright. Welcome, everyone, and thank you very much for joining us today. I'm Robert Hilson of Logical, and I'm really excited about what we're sharing with you this afternoon. We're gonna introduce something called Ask, which some of you have heard about. It is a new AI powered search data synthesis and data analysis tool that we pioneered in Reveal, and now it's available in Logical in the EMEA region. We're really excited about it. Before we dive in, I wanna introduce my colleague, Andy Punter, renowned discovery consultant, Logical specialist, international man of mystery. Andy is a sales engineer for Reveal, a senior sales engineer for Reveal and has had more frontline conversations about us than just about anybody in the company. So, Andy, we really appreciate you you being here today. Thanks, Robbie. Really excited to be here. Right on. So let's let's quickly go through the agenda. We're gonna try to keep this to thirty minutes, a a tight thirty minutes. So we're gonna be moving pretty quickly. We're gonna start with how ask kind of actually works under the hood, not so much from a technical perspective, but functionally so you understand, like, what's happening when you ask a question. Then we'll look at where it fits into your actual workflow all the way from, you know, early case assessment to starting to, you know, identify key facts around your case through to deposition prep, trial prep, etcetera. And then we'll actually, get into a deep dive demo, and finally round out with how you actually get access to this thing. So before we get into how it actually, works as I go to the next slide, let's touch on briefly, like, what Ask is. It is a, data search and data analysis and synthesis tool that, allows you to pose natural language questions about your entire document set or subsets of documents and get back cited answers that help you actually understand what's happening in your matter if you need to come up to speak quickly on things like key custodians, timelines, events, key facts, etcetera. So the core of this is is, I think, pretty simple, and and that's what makes it powerful. It searches your documents as searches your documents semantically, meaning it's looking for, you know, beyond just keywords, meaning and intent. And then what it's gonna do is actually surface relevant materials and synthesize them into clear, readable answers. One of the critiques that, you know, we we hear of some of the AI tools in the market that we're we're certainly mindful of as we're not only building these tools, but bringing them to market is that they're they're not transparent. Right? Meaning, it's hard to know kind of how the AI is deriving its answers and decisions, whether that's answering questions like Ask does or making coding decisions like with our, you know, with our Agi solution. So, critically and and Andy's gonna show you this. But Ask is built such that every answer comes with citations and links back to the actual source documentation. So you're never you're never just trusting the AI. You can actually, verify, the answers that you're getting back, and we'll, again, kinda show you how that works. So before we actually get into the demo, let's just kinda briefly touch on, you know, kinda where you can get the most use out of this. And and, Andy, I'll, I'll kick it over to you in a second. But what I find compelling about us is that, you know, it's not it's not just a point solution for, like, one part of the process. It's really gonna fit across the entire life cycle of a matter. So, you know, on day one, before you even review a doc, you can start asking questions and building case strategy based on, you know, what's actually in the documents. So, you know, what what might that look like? During early case assessment, for instance, instead of starting with, you know, your your typical kind of logical filtering, keyword searching, etcetera, you're getting answers to natural language questions in seconds, and that's that's actually informing your approach. Andy, I'll ask you to to jump in here to a a demo here in a second. But, just from your perspective and kinda based on what you're hearing from kind of early users, like, where are they getting kind of the most value across the across the process? That's a great question. I think we're starting to see a huge change from that. What have I got? Right? Where have I got something? Where can I find it? And if we hark back, I don't know, five, two years ago, that would have been keyword searching, trial and error review, back to keyword searching, date range filtering, custodian filtering. And I'll still have to look at samples of documents to find the knowledge that's gonna go forward in my investigation to build my case strategy, to build even my investigation strategy and review strategy. Now we're seeing that time drastically being reduced all through the power of ask. And I can ask questions, and we'll see this in a in a short amount of time when I can dive really deeply so quickly just by knowledge I'm getting from from documents and then pull back out and go to another one. We're seeing this used in triage, triage and investigations as they come in. Do the investigations have merit? Do we have the right type of data? We're seeing in all the way through to QC. Have I put coded and reviewed all the correct documents? Have I reviewed them correctly? Am I gonna have some outliers that I need to look into further? And we're seeing what I'm enjoying most is we've released this with what we had in mind. But seeing how clients are going and finding intuitive ways to use Ask, Yep. and we're finding new best practices, and we're finding use cases we didn't even think of. That's the most exciting part of this. Yes. And what's interesting to me is like, it's not just that this usage is happening across like various phases of the but it's happening for different use cases as well. So you mentioned triage of investigations and litigation, it's actually you know, people are finding uses for, you know, DSAR, subpoena response, uses like that as well. Yeah. Absolutely. It's it's pretty cool. Awesome. I'm gonna, because you reminded me, I'm gonna stop sharing my screen. We can get into into the demo. Before we get into this, I do wanna encourage people to ask to ask questions. As they arise, we'll get to as many as we can. We did this a couple days ago and had had, I think, 30 or 40 questions. So please fire away, and and we'll get them to Andy. Andy, I'll stop sharing my screen. Over to you. Sorry. Alright. Let me share our screen. Amazing. Cool. So what I'm gonna talk you through is two use cases that we're seeing Logical and and especially Ask on top of Logical start to to really define and help. So the first is gonna be an introduction to us, but also an introduction to a case. Hey. I'm a brand new lawyer, and I have just been thrown headfirst into this matter, and I know absolutely nothing. So I need to have a great place to go and find knowledge about that matter. So what we're gonna do is within the logical UI, you will know that the search bar at top at the top here is where I can do my searching. I can do my keyword searching. And what's more, I can now ask AI. So I can use ask right in my searching. Because as Robbie mentioned at the start, this is all about the next level of searching. So I'm gonna start with a basic question. For example, who was the CFO at Enron? Now this, I know, is a pretty basic question, but it really gets you going in terms of where it is. Because what's happening right now is I'm generating a response and hey. That was incredibly quickly. I'm generating a response based upon my document set, based upon these 38,000 documents that I'm coming into. Not going to Facebook, Wikipedia, LinkedIn, anywhere else. I'm just basing it upon your documents. And, hey, let's look at the answer. Right? Andrew Faster was CFO and on before being replaced by Jeff Mulhoney. So before I even look into the context and what that actually means, I immediately have two people of interest from one question. K? And we all know know you guys listening in need evidence to prove to prove this information, and we're gonna give you that from the forms of your documents. If I click on what this number one, that's going directly to a citation within your document. Click on number two. Again, another sub subset of that single document to show you how we're getting that information because this is really, really important is, yeah, I found this really incredible document. What does it actually mean? Where did it come from? Okay. It actually came from discovery. Actually came from your document pool. And, again, we're gonna give you up to a 100 other references, which you can all see down here. These are actual documents in this matter that help me gonna provide evidence to why I've got this knowledge, which is really, really cool. Andy, let me pause you there. Like, as we're surfacing these. other up to a 100 kind of semantically similar documents, can you speak to, like, how you can actually use those results to kind of take an iterative approach to asking questions and digging deeper into your document set? Yeah. I think that's a really, really good point. So, for example, I can see my first document here, which means that I can now start reading that snippet. I can start thinking about reasons as to why or what I might want to ask my next question. I can review. I can pop open these documents, read them in more context. I can start tagging, and then I can start thinking about where do I want to go next. Now if I'm looking at this investigation from the front, I want to know why Andrew Frastal was replaced. I want to know anything around this kind of area, and maybe that's just where I might need a a a new bit a bit more information because I can see within here, Andrew Fassow was put on a leave of absence around the mine period. Why did that happen? And I can continue to review. I can also see some interesting things around investment partnerships. Now this is a really crucial part of the Enron story. But if I don't know that, which as a brand new lawyer on this case, I don't know this currently right now, I can do some pretty impressive things with this because, hey. I don't know about it. Let's ask a follow-up question. What was LJM? And I'm getting that information from this exact document here. Right? What was it? So rather than just repeating my question, I'm gonna do something slightly different, which is also incredibly cool. You'll notice I've only been using one line of this text down below. Now below that, there is a second line. This additional instructions is a really powerful piece of technology that gives me the ability to tell ask how I want my content back to me. So, for example, I'm gonna say answer in timeline. And this is gonna provide me a a really nice timeline of the thing I've asked based upon the documents that are being brought back. Now all caveat, this is not a timeline that many of you guys would want from your general counsel or outside counsel or would be able to provide yourself. But the thing you're gonna provide is after two weeks of review, a lot of coffee, and a very large Excel, what I'm gonna give you is a really cool starting point for key things that happened in key dates so that I can now target specific document sets. Look. I can now target the origin of LJM. I can now target its business relationships and its its its reasons it's it bought capital from Enron. These 24 business relationships, that's something I really wanna target and really wanna investigate in. So from here, I can now see how easy it is to start diving down rabbit holes. And, again, like we said, we're giving you those one supporting documents so that you can confirm the answer you're getting is correct, which, is really, really powerful. Andy, let me ask you this. And I I wanna say that I I heard this from from one of your colleagues on the on the team. But in thinking about kinda how this can work from an early case assessment standpoint and kinda doing that, like, really kinda that that first kind of run through of your documents and thinking about the tagging structures and everything. Like, I imagine, if if you're kind of searching for for issues or people like your, you know, your your CFO, your and, CFO example, Yeah. Is there is there an option just to kinda tag everything and say, you know, these it surfaces 61. These are probably kind of relevant to the same search that I'm running, and then kinda use that as a kind of a deeper dive where you can start using the calling filters to get to get deeper into the data? Yeah. I mean, the beauty about how we've engineered Ask in both logical and Reveal is how integrated this is. Right? So exactly as what I mentioned. Right? Clicking on these more docs, open those 61 supporting documents, and in the background, you'll now see those 61 supporting documents ready to review, ready. to tag, ready to bulk, and ready to organize. Right? I can now go, okay. Team member number one, step up. I want you to review those 61 documents and really provide that that interest as well. Now what I can also do is now take that a next step further. If I go back into ask and go back into my previous session, which, again, a really nice feature. We have that short term and long term memory. So I can go back into a repeated question, and I can see where I got that answer from yesterday before I took a coffee break, and it disappeared. It's never got it's never truly gone as well. I really wanna know these 21 business 24 business relationships. So I can say what transactions now those eagle eyed amongst you would also notice my my severe dyslexia has also thrown in a spelling mistake in occurred. Now this is also one thing I wanna wanna mention is this we all ask questions differently, whether that is from where you're from, your education, your background, or just how you speak. Ask interprets the question and actually understands the meaning behind it. So I can add spelling mistakes. I can add for example, you can see my name here, Andrew Punter. I may go as Andy. I may go as Drew. Don't go as Drew. But I could go as multiple different names, and we're gonna use that and collate that into what is actually relevant so you can ask questions in different ways, but still get the answer you need. So I'm also gonna add answer this in a timeline because I want to see when transactions occurred, when things happened, when that happened within my dataset, and it's gonna get me back to that. So the key thing about how we ask, how we talk is this fundamental meaning behind it, and that's what ASCII is gonna give you depending if I ask the question or if yourself, rather, you ask the question in a slightly different way. And so, Andy, I mean, this is gonna be, like, intuitively kind of an intelligent search, I guess. But even, like, given what you said, are there are there kind of best practices in terms of, you know, making sure that you're kinda putting forth the best prompt and you're getting back the the kind of the best response? Yeah. I that's a really, really good question, and it all comes down to the usability of Ask. So what I like to to talk about is so first of all, I'm just gonna say this is one of my favorite answers that comes from this. Like, this is now really deep into where I want to be for this investigation. But in terms of us, now for investigations, what I like to think of for investigators out there and and and lawyers is you are asked to provide arguments to or from a very large holistic question. Right? So let's take the idea of fraud. If I write in, was there any fraud? Unless someone has gone, thanks, Robbie. I had a great meeting. I'm off to go commit some fraud now. It's it's not gonna give you the right answer. But you and the reason you are an investigator, we still have lawyers, and we still have amazing people doing these things, is you break that up into second level, third level, fourth level, fifth level questions to sum up into that argument. And as you get further down that tree, those questions become more and more fact driven as opposed to subjective. And that's those questions, those third, fourth, and fifth level questions, that's where ask really helps you surface the facts that ultimately go all the way up into your argument that you may be presenting an internal investigation, arbitration, litigation, anything where you need to find evidence. And that's what Ask really does. Excellent. Let me let me ask you this, and this is a a common question when we did this a couple days ago. But what is, like, what is the universe that this might be an obvious question, but, like, what is the universe of content that Ask is actually pulling from? I would assume that it's just text in in documents that is, you know, that is actually processed by Logical. Is there I mean, is that is that kind of the scope? Yeah. That's absolutely right. So it's all the text that's gonna be processed. So that is the body of an email. It's the document itself, the words within a Word document, those kind of things. So all of that all that messaging that's been brought in that you can search your keywords over, that's what we're we're bringing across. So so we're not gonna be able to do things like metadata searches, but it yet. will pull in, yes. Not good point. Not yet. But it will pull in things like, you know, embedded text that's associated with images. And, because Logical is, it's it's OCR ing hand handwritten documents, for instance, is gonna pull in that that type of information. as well. Absolutely. That's a really good point. So think of those construction matters where I have lots of text within documents, within pictures. We're pulling that, and we're going through OCR. We're going through deep text recognition, and all that text is now searchable and analyzed. But your point to metadata. Right? If I look back at Logical, I have a great tool to search metadata. I can filter down via my conversations, via my chat channels using our easy filter wheel, then ask questions on that population. So, hey. Let's just filter between myself and Robbie's emails here, and then I'll start asking questions. One thing I do wanna talk talk about earlier, is earlier you mentioned about, hey. Lots of people talking about keywords and stuff like that. Now if I have this as a as an idea, keywords we all know are very, very good, but they are very targeted. So for example, if I've just come back off an interview and something has been a flipping comment has been talked about about this project. So for this, I'm gonna say the project is called the EPE project. And if I run a search term for that, I don't get any results in my population, which is a major problem because now I may have to go back to the person, back to the deposition, back to the interview, and identify what okay. And clarify what they meant. Well, what if I just asked ask? So what was the EPE project? And this shows you the kind of flexibility that comes from the truth within your documents. And when I was talking earlier about how we ask questions and what ask is interpreting, I can now start finding meaning from phrases or concepts as opposed to the exact phrase itself. Because I can see, for example, the documents indicate that the EPE was a Brazilian partnership with an energy company. Okay? Mhmm. Someone may have miss misspoken, and I can follow-up. Okay? Okay. That's that's really cool. But I can do things such as were there any additional other other names to this project? And okay. Let's just ask that. Right? Do I know about this project? Do I know about this this thing? What else could this thing be called? And, again, the classic always do stuff on live data. And, again, if I don't get a result, I can just ask my question again. And what other names were there for the EPE project? I mean, so you can really see how this, like this is going to inform your keyword approach if if that is the tactic that you were taking. Yeah. Absolutely. And I can I can as as you see here, I can just rephrase that question to get a result or not in this case? Let me ask you this, and you you began to speak to this, but it's always interesting interesting to me in talking to, to Logical customers and depending on kinda where they live. So if they're, you know, if they're in house versus they work on the the law firm or advisory side or or government, kind of where they start their approach to calling documents to getting into the dataset. Do you find that it's more useful to, for instance, start with ask to start to kind of identify these these key things or to start to, for instance, apply some of these filters and then add ask on after the fact. I think there are pros and cons to both ways. And it's a very sit on the fence answer, but there are cases where you just need to filter the data because you know you need to filter down and just provide. Let's take a DSAR, for example. A DSAR is very different to an internal investigation. I just need to respond to the request. Now what we're seeing is DSARs being weaponized more and more so that they are now going into employment tribunal, but now the data subject or the the ex employee, the disgruntled employee has a huge amount of data before your employment lawyers have even looked at the data. Yep. So how do I get up to speed? I can use ask to really work out what has been sent, what else needs to be, provided. Is there anything else that needs, to help with my defense, as as a company and make sure I have my facts and and everything completely straight. And that is one of the best ways we can see us being used, especially over over here in this side of the world on those kind of really complex matters as well. Yeah. That's awesome. Let me ask you guys, and this this question came up the other day as well. How does ask and I'm I'm asking in the in the context of, like, an HR investigation, for instance. How. does it handle things like like sensitive or or inappropriate conduct? Like, that that type of language. I I guess kind of a two part question. And then kind of types of data that most often lend itself to that type of communication, for instance, Slack Slack data. Yeah. It's really good. So I I I can as we bring data into Logical, especially for for chat messaging, we're gonna break that up into twenty four hour chunks. So there's gonna be a decent amount of context left and right of my potentially person point. But, also, as we bring data into us and we search across it, we're gonna say we're searching for those really important paragraphs, sentences within your document for a multiple number of reasons, but it also brings you right to the message that could be inflammatory. It could be derogatory. It could be really just just plain nasty. And all of those things help surface and bubble those up to the top as well, especially using ask. Excellent. Good stuff. So, anyway, I wanna be mindful of people's time. We have about five minutes left. Is there is there anything else that you wanna show? Not today, but I would love to show anything towards your use case if you just as as Robby mentioned at the end, how to get in contact with us. Yeah. Fantastic. So we'll hit that briefly. I just wanna kick share kind of a couple more things around kind of best practices here. Talk about do's and don'ts. I guess the short version of this would be, you know and, Andy, you touched on this, but, like, be as specific as possible. Try to keep your questions focused. And then, you know, if if it makes sense to, like, use these filters before you run out so that you're you're you're you're running against a narrow kind of slice of the data. I think, I wanna highlight on the the what to avoid, and, Andy, I'll I'll toss it to you and and you tell me your, your don'ts. But, you know, there's I think there's a temptation to ask kind of, like, sweeping questions, like, tell me everything about the company's finances as opposed to, you know, some of the more targeted questions that that you ask. The the more specific questions are are typically going to highlight more valuable information that you can kinda feedback into your your approach. Ask is a it's a precision tool, and you're gonna get the most out of it if you if you treat it that way. Andy, anything that that's kinda that stands out to you in terms of either best practices or, you know, things to avoid? I think that's absolutely right. Specificity is king. The other thing is don't use us for for the way of searching we already have within the tool. Right? We already have fantastic communications filters. We have great date filters. We have those. Use those in conjunction with us to find your population and then ask questions as opposed to getting asked to find the small subject to then ask the question. Excellent. So why don't we why don't we wrap up with this important question? Like, how do I actually get access to ask? There are really two ways. One, if you are not an existing customer of Logical, you can become a customer. We offer this standard in our subscription packages. And, basically, the way that it works is it's gonna be a small kind of premium to your monthly storage rate, and that gets you a bucket of questions per gigabyte. So, you know, the the more data volume that you have in your storage plan, the more ask you get. We are and this is important for people that are existing customers. We're offering a free trial of this thing, essentially. So the way it works is if you are a subscription customer and you wanna use this on an existing matter, you can basically pick any case that's under 250 gigabytes, and we're gonna give you 250 free questions to do with what you will. If you wanna use this on a net new matter, and we'd encourage you to do that, you can use it on anything that's under 500 gigs. And for those, we're gonna give you up to 500 questions to to use on your matter. So if you're interested in taking advantage of this, especially the the the free trial offer, which is it's gonna be with us for a limited time, so I'd I'd invite you guys to jump on it. You can reach out to your account rep or you could just go to the website, go to our our ask page, which is logical.ai or logical.i. Request, access, and we'll we'll make sure that you, you get in touch with the right people. Speaking of, if you have any follow-up questions to this or, again, if you just you need kind of a direction as to who do I reach out to you to actually get access to ask, feel free to reach out to either of us. We'll we'll get back to you, as soon as we can. Andy, it's always a pleasure. Thank you very much for doing this. Any any, kinda closing words from you? Nothing, but I'm really excited to speak to every single one of you if you can, about how you want to use Ask, and it'd be amazing to talk about. Yeah. Same here. And, thank you all very much for for joining us. We know you're you're busy, we appreciate you making time. We'll see you next time. Thanks all. Thanks, Robbie. Bye. Bye.